Health Tech News: AI Scribes Scrutinized, Telehealth Under the Microscope, and AI in Drug Development
This week’s health tech news focuses on the impacts and implications of AI in healthcare.
AI Scribes: Promise vs. Reality
The Peterson Health Technology Institute (PHTI) released a report assessing health systems’ experiences with AI ambient scribes. The report highlights that while these tools are marketed as time-savers, evidence of their effectiveness is limited. The PHTI also explores how AI scribes may be contributing to higher healthcare costs.
According to the PHTI, the industry is still determining how best to evaluate these systems. One of the key challenges ahead will be to develop and implement standardized methodologies across different indicators to better understand the overall impact of these tools and their long-term effects on efficiency.
Questionable Telehealth Partnerships
Drug manufacturers like Pfizer and Eli Lilly are increasingly partnering with telehealth platforms to deliver treatments to patients. This practice is drawing scrutiny from senators concerned about potential violations of federal anti-kickback statutes.
Researchers Ateev Mehrotra, Olivier Wouters, and Erin Fuse Brown at Brown University have been studying this market. Their findings highlight the possibility these partnerships might improve access to care, but also potentially lead to inappropriate prescribing. The researchers note that assessing the impact of these partnerships is challenging without greater transparency.
Breakthrough Devices and AI Advancements
The FDA has granted “breakthrough” status to numerous experimental medical devices, including PathChat DX, a generative AI chatbot from Modella AI developed with Mass General Brigham. This device helps pathologists in diagnosing cases. Other advancements include a breast cancer test, a breast cancer imaging tool, and two Alzheimer’s blood tests.
AI in Drug Development
Apheris, a life sciences data company, announced its consortium of drugmakers will contribute to OpenFold3, an open-source protein-folding model based at Columbia University to accelerate drug discovery. Additionally, AbbVie and Johnson & Johnson will contribute structural data to the project.
At a recent conference, AstraZeneca presented a report on its use of an AI model to analyze CT scans and predict survival in lung cancer patients. Developed by Altis Labs, this technology is also being tested for breast and colorectal cancer scans. Google has also released TxGemma, large language models tailored for drug development, based on the work of Google DeepMind.
